Magnetic flux sensor and method

Abstract

A device comprises a first magnetically conductive structure, a second magnetically conductive structure, and a circuit coupled the first and second magnetically conductive structures. The circuit senses information pertaining to a permeability of the first magnetically conductive structure relative to a permeability of the second magnetically conductive structure. The device may be used for various flux sensing and control applications, such as in current sensors.

Description

CROSS-REFERENCE TO RELATED PATENT APPLICATIONS

This application is a Continuation of application Ser. No. 10/251,383 filed Sep. 20, 2002 now U.S. Pat. No. 6,774,618.

This application is a continuation-in-part of U.S. application Ser. No. 09/459,683, filed Dec. 13, 1999, now U.S. Pat. No. 6,456,059, incorporated herein by reference in its entirety.

BACKGROUND OF THE INVENTION

1. Field of the Invention

This invention relates to magnetic flux sensors and methods. This invention also relates to sensor methods and systems that utilize magnetic flux sensors to acquire information pertinent to another ultimate parameter of interest, such as current.

2. Description of Related Art

A common problem with magnetic devices is that there is often no practical way of knowing how much magnetic flux is present in the device. This parameter is of obvious interest in any magnetics application, but is of particular interest in applications-where the magnetic material that carries the magnetic flux is liable to become saturated.

Devices that in some way utilize magnetic flux are common and have been employed in a diverse array of applications. For example, magnetic devices such as transformers are commonly used by utilities and in various household and industrial applications to convert power sources from one voltage level to another voltage level. Another type of transformer is a current transformer, which is a widely employed device for performing current measurements. Magnetic devices also include electromechanical devices such as relays, electromagnetic contactors, electric motors, and electric generators. Relays and electromagnetic contactors are used to control whether a particular electrical connection is opened or closed. Electric motors and electric generators are used to convert electrical power into mechanical power and vice versa. Numerous other magnetic devices also exist.

Current transformers provide an especially good example of the problem. A current transformer comprises primary and secondary windings that are wound about a transformer core. A primary current I₁ flows through the primary winding and induces a magnetic flux which flows through the transformer core. The magnetic flux in turn induces a secondary current I₂ in the secondary winding. For a linear (unsaturated) current transformer, the primary current I₁ is related to the secondary current I₂ by the following relationship:
I₁N₁=I₂N₂  (1)

Therefore, since the parameters N₁ and N₂ are known (N₁ and N₂ are the number of turns of the primary and secondary windings, respectively), the primary current I₁ may be measured indirectly by measuring the secondary current I₂. The secondary current I₂ may be measured by placing a burden resistor across the secondary winding, and measuring a voltage V₂ developed across the burden resistor as a result of the secondary current I₂: $\begin{array}{cc}{I}_1=\frac{V_2}{R_b}\left(\frac{N_2}{N_1}\right)& \left(2\right)\end{array}$
where R_b is the resistance of the burden resistor. In short, therefore, the primary current I₁ may be determined by measuring the voltage V₂.

The voltage V₂ that is developed across the secondary winding is related to the net magnetic flux Φ in the following manner: $\begin{array}{cc}{V}_2=N_2\frac{d\Phi }{dt}& \left(3\right)\end{array}$
(Although Eq. (3) is sometimes written with a minus sign before the right-hand term, a minus sign is not used herein.) In saturation, since the net magnetic flux Φ stays at a constant saturated level, there are no time varying changes in the net magnetic flux Φ. As a result, there is no voltage developed across the secondary winding (V₂=0) and there is no secondary current that flows through the secondary winding (I₂=0). This is true even though current continues to flow in the primary winding (I₁≠0). It is therefore apparent that, in saturation, Eqs. (1) and (2) do not apply and the primary current I₁ cannot be measured.

Typically, saturation can be avoided by only measuring currents above a certain frequency and below a certain magnitude, these operational limits being determined by the construction of the current transformer. However, sometimes low frequency components appear unexpectedly in the primary current, causing the current transformer to go into saturation. Therefore, knowing the amount of magnetic flux in the transformer core would be highly advantageous, because it would provide an opportunity to take measures to counteract the low frequency components that would otherwise cause the transformer core to saturate. Indeed, it would be even more advantageous if those low frequency could not only be counteracted, but measured as well.

This same general phenomenon also exists with respect to other magnetic devices. For example, synchronous electric motors operate through the creation of a magnetic field that rotates in synchronism with the rotor. The rotating magnetic field is generated by providing the stator with sinusoidal drive current. However, given that the drive current is often electronically-generated, it is possible that DC and/or other low frequency current components can “creep into” the drive current, causing the magnetic material in the motor to tend toward saturation. Such current components can occur, for example, if the switching transistors used to generate the sinusoidal excitation current are not perfectly matched. Magnetic losses in motors often help avoid saturation, but low frequency current components nevertheless at least cause the motor to operate less efficiently.

Even ignoring the problem of saturation and low frequency current components, it is often desirable for other reasons to monitor the magnetic flux in a magnetic device. The provision of a rotating magnetic field is a fundamental aspect motor control. Typically, however, the magnetic flux in the motor is not directly measured but rather is assumed to have a certain value (or distribution of values) based on the known current that is applied to the motor. Being able to directly measure the magnetic flux in an electric motor would provide an opportunity for better, more-efficient control of the motor.

Likewise, for electromagnetic contactors, or for other devices in which an electromagnetic field provides an actuating force for moving a mechanical substructure, a direct measure of the magnetic flux would allow the actuating motion to be controlled more precisely. This could be used to improve operation of the device or to effect other desirable results, such as extending the life of the device.

Magnetic flux sensors have previously been provided. For example, current transformers are one type of magnetic flux sensor, i.e., because a current transformer operates by having a secondary winding that senses magnetic flux in the core of the transformer. Conventional current transformers, however, are not well-suited to measuring low frequency flux components for the reasons previously described.

Another type of magnetic flux sensor is the Hall-effect sensor. When a conductor carrying a current is placed in a magnetic field, a voltage is created across the conductor in a direction that is perpendicular to both the direction of the magnetic field and the direction of current flow. This well known phenomenon is referred to as the “Hall-effect,” and is the operating principle for Hall-effect sensors. Magnetic flux sensors that operate based on the Hall-effect have been employed in a diverse array of applications, such as current sensors.

A primary disadvantage of Hall-effect sensors, however, is that they must be placed in the magnetic path, which usually requires that a gap be made in the flux-carrying material. Given the extremely low permeability of air (approximately 1.0) as compared to most core materials (in the range of 104 to 105 depending on the material used and operating conditions), the insertion of an air gap, however small, has a dramatic and usually undesirable effect on the magnetic characteristics of the system. For example, conventional current sensors that use Hall-effect devices have significantly poorer resolution and accuracy than current transformers over those operating ranges in which current transformers do not saturate.

Therefore, what is needed is an improved method and system for magnetic flux sensing. What is also needed is an improved method and system for flux sensing that is capable of operating in the presence of low frequency flux components, and even more preferably capable of measuring those low frequency flux components.

BRIEF SUMMARY OF THE INVENTION

In accordance with a first preferred embodiment of the invention, a device comprises a first magnetically conductive structure, a second magnetically conductive structure, and a circuit coupled the first and second magnetically conductive structures. The circuit senses information pertaining to a permeability of the first magnetically conductive structure relative to a permeability of the second magnetically conductive structure.

In accordance with another preferred embodiment of the invention, a method of sensing a parameter comprises acquiring information indicative of a permeability of a first material relative to a permeability of a second material. The permeability of the first material relative to the permeability of the second material is indicative of the parameter.

In accordance with another preferred embodiment of the invention, a system comprises a magnetically first magnetically conductive structure formed of a first material, a second magnetically conductive structure formed of a second material, first, second and third windings, and processing circuitry. The second material has a different permeability than the first material. The first winding is wound around the first magnetically conductive structure and not the second magnetically conductive structure. The second winding is wound around the second magnetically conductive structure and not the first magnetically conductive structure. The third winding is wound around both the first magnetically conductive structure and the second magnetically conductive structure. The processing circuitry is coupled to the first, second and third windings and receives information pertaining to a permeability of the first material relative to a permeability of the second material. The processing circuitry generates information pertaining to a parameter of interest based upon the relative permeability of the first and second materials.

In accordance with another preferred embodiment of the invention, a system comprising a magnetically conductive structure and a feedback control circuit. The feedback control circuit further comprises a sense winding and an excitation winding. The sense and excitation windings are wound around the magnetically conductive structure. The feedback control circuit acquires feedback information pertaining to a voltage across the sense winding and provides an excitation signal to the excitation winding to control the amount of flux in the magnetically conductive structure.

Other objects, features, and advantages of the present invention will become apparent to those skilled in the art from the following detailed description and accompanying drawings. It should be understood, however, that the detailed description and specific examples, while indicating preferred embodiments of the present invention, are given by way of illustration and not limitation. Many modifications and changes within the scope of the present invention may be made without departing from the spirit thereof, and the invention includes all such modifications.

BRIEF DESCRIPTION OF THE DRAWINGS

A preferred exemplary embodiment of the invention is illustrated in the accompanying drawings in which like reference numerals represent like parts throughout, and in which:

FIG. 1 is a flux sensor in accordance with a first preferred embodiment of the invention;

FIGS. 2A and 2B show the construction of a transformer for the flux sensor of FIG. 1;

FIG. 3 is a graph of the permeability of two materials used to construct the transformer of FIGS. 2A and 2B;

FIG. 4 is a graph showing the reluctance of two materials for a transformer of the type shown in FIGS. 2A and 2B;

FIG. 5 is a graph of magnetic flux for two core materials, as well as a total magnetic flux, as a function of a voltage sensed across a sense winding in the system of FIG. 1,

FIG. 6 is a graph of magnetic flux density for two core materials as a function of a voltage sensed across a sense winding in the system of FIG. 1;

FIG. 7 is a graph of magnetic field intensity for two core materials as a function of a voltage sensed across a sense winding in the system of FIG. 1;

FIGS. 8 and 9 are graphs showing the operation of the system of FIG. 1;

FIG. 10 is a current sensor that employs the flux sensing system of FIG. 1;

FIG. 11 is another current sensor that employs the flux sensing system of FIG. 1;

FIG. 12 is another current sensor that employs the flux sensing system of FIG. 1;

FIG. 13 shows an error transfer function for the system of FIG. 12;

FIG. 14 is one possible implementation of the system of FIG. 12;

FIG. 15 shows a flux feedback control loop; and

FIGS. 16A-16B are flux feedback control loops that comprise a voltage regulator circuit.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

Referring first to FIG. 1, a flux sensing system 10 in accordance with a first preferred embodiment of the invention is illustrated. By way of overview, the system 10 includes a transformer 14, flux sense circuitry 16, and a high frequency excitation source 18. The transformer 14 comprises first and second transformer cores 20a and 20b, a sense winding 22 which is associated with the flux sense circuitry 16, and an excitation winding 24 which is associated with the high frequency excitation source 18. The system 10 is used to measure flux in the transformer cores 20a and 20b. In FIG. 1, the flux is induced by current flow in a winding 26 coupled to a current source 28. Of course, the flux could be developed in the transformer cores 20a and 20b by some other mechanism. Additionally, while the magnetic device is shown in FIG. 1 to be a transformer, the invention is equally applicable to other types of magnetic devices.

FIGS. 2A and 2B show the construction of the transformer 14 in greater detail. As previously noted, the transformer 14 is constructed of first and second transformer cores 20a and 20b. In order to facilitate construction, the transformer cores 20a and 20b preferably have the same mean path length (L_e1=L_e2), although transformer cores with different mean path lengths could also be used. To fabricate the transformer 14, the transformer core 20b is first wound with the sense winding 22. Then, the transformer core 20b and sense winding 22 are placed adjacent the transformer core 20a (FIG. 2A), and the combination of these structures is wound with the excitation winding 24 and the primary winding 26 (FIG. 2B). The sense winding 22 is therefore wound only around the transformer core 20b, and not the transformer core 20a, and the excitation winding 24 and the primary winding 26 are wound around both the transformer core 20 and the transformer core 20b. This is shown schematically in FIG. 1 by virtue of the core 20a being shown not to extend adjacent the winding 22. The sense winding 22 and the excitation winding 24 may each, for example, comprise between 100 and 10,000 turns, although other number of turns could also be used. The primary winding may comprise a smaller number of turns, for example, one turn.

The transformer cores 20a and 20b are constructed of materials with different magnetic characteristics. In one embodiment, the material for the transformer core 20a (or analogous component in most other flux sensing applications) is Supermalloy provided in the form of a tape wound core. This material can be purchased as Part #343 P 4902 from Magnetic Metals Corporation, 2475 LaPalma Avenue, Anaheim, Calif., 92801. For flux sensing applications in which the ultimate parameter of interest is current, the core 20a may for example be formed of Supermalloy, Part #343 P 8602, available from the same company. The transformer core 20b may for example be formed of Permalloy 80 provided in the form of a stamped lamination. This material can be purchased as Part #R-510-14D from Magnetics Inc., 796 East Butler Road P.O. Box 391 Butler, Pa. 16003.

FIG. 3 shows permeability functions 30a and 30b for Supermalloy and Permalloy 80, respectively. The permeability functions are both a function of flux density, but the permeability functions vary with respect to each other in a way that is not divisible by a single scalar constant. Thus, although both functions vary with respect to flux density, they also vary with respect to each other in a manner that varies in accordance with flux density. Since magnetic flux density is related to magnetic flux and-magnetic field intensity, the permeabilities of the transformer cores 20a and 20b also vary with respect to each other as a function of these parameters as well.

FIG. 4 shows this effect more dramatically. FIG. 4 is a graph of the reluctance of the transformer core 20a (curve 32) with respect to magnetic flux, the reluctance of the transformer core 20b (curve 34) with respect to magnetic flux, and the total reluctance of the transformer 14 (curve 36) with respect to magnetic flux. The magnetic properties shown in FIG. 4 have been exaggerated somewhat for purposes of explanation. The curves 32, 34 and 36 do not accurately describe the preferred Permalloy and Supermalloy 80 materials.

The transformer cores 20a and 20b serve as a flux divider and carry varying proportions of magnetic flux in accordance with the total amount of flux flowing through the transformer 14. In particular, reluctance is equal to the mean path length L_e divided by the product of the effective area A_e and the permeability $\left(ℛ=\frac{L_e}{A_e{\mu }_r{\mu }_0}\right).$
It is therefore seen that, in operation, as the proportion of μ₁ to μ₂ varies, the proportion of the magnetic flux that flows in the transformer core 20a relative to the proportion that flows in the transformer core 20b will also vary. For example, if $\frac{{\mu }_1}{{\mu }_2}$
is decreasing and therefore $\frac{{ℛ}_1}{{ℛ}_2}$
is increasing as flux increases (although the opposite approach could also be used), then the reluctance R₂ for the transformer core 20b will be decreasing relative to the reluctance R₁ for the transformer core 20a, and a greater proportion of the flux will flow in the transformer core 20b than in the transformer core 20a. As described in greater detail below, the excitation signal applied to the excitation winding 24 causes high frequency flux excursions that are experienced by both the transformer core 20a and the transformer core 20b. However, the decreasing reluctance of the transformer core 20b relative to the reluctance of the transformer core 20a causes the portion of the flux excursion experienced by the transformer core 20b to increase. In turn, this causes the peak-to-peak voltage of the signal appearing at the sense winding 22 to experience a corresponding increase. The voltage across the sense winding 22 can therefore be monitored to obtain an indication of $\frac{{\mu }_1}{{\mu }_2}.$
Since $\frac{{\mu }_1}{{\mu }_2}$
is a function of the amount of flux in the transformer cores 20a and 20b, obtaining an indication of $\frac{{\mu }_1}{{\mu }_2}$
also means that an indication of the flux in the transformer cores 20a and 20b is obtained.

Referring back to FIG. 1, in practice, the flux sense circuitry 16 preferably comprises a high impedance voltage sensing circuit. A high impedance circuit is preferred so that practically no current is induced in the sense circuitry 16 during operation of the system 10. In practice, the flux sense circuitry 16 could be simply an operation amplifier connected to sense the voltage across the sense winding 22. In a more elaborate embodiment, the flux sense circuitry could be an A/D converter connected to digitize the voltage across the sense winding 22. As will become apparent below, the voltage across the sense winding 22 is indicative of the magnetic flux density in the cores 20a and 20b. However, because this relationship is not linear, it may also be desirable to linearize the output of the flux sense circuitry. In this event, the output of the A/D converter may be processed by a microprocessor that has access to a look-up table to perform such a linearization. Additionally, if it is desired to know an actual value of the magnetic flux density, rather than just whether the flux is increasing or decreasing or staying constant, then the flux sense circuitry may also comprise a microprocessor that has access to a look-up table to convert the sensed voltage into a flux density value. The manner in which the sensed voltage can be converted to a flux density is described in greater detail below.

The high frequency excitation source 18 supplies a high frequency excitation signal to the excitation winding 24. The frequency of the excitation signal is high relative to the frequency of the current in the primary winding 26, in order to permit the portion of the voltage appearing at the sense winding 22 attributable to the excitation applied at the excitation winding 24 to be distinguishable from the portion attributable to the excitation (i.e., the primary current) applied at the primary winding 26. The excitation signal may for example be a 20 kHz square wave voltage excitation signal. However, the optimal frequency of the source 18 may be different on the anticipated frequency of the current source 26, and depending on the technique used to detect the polarity of the magnetic flux, as described below. Additionally, although the mathematical description given below assumes a square wave voltage excitation signal, other types of signals could also be used.

The operation of the system 10 will now be described. The mathematical derivation that follows is given for purposes of explaining why the voltage across the sense winding 22 is indicative of the magnetic flux, magnetic flux density and magnetic field intensity in the transformer cores 20a and 20b. As will become, apparent, it is not necessary that any of the following equations be calculated during the operation of the flux sensing system 10.

In operation, a primary current is applied at the primary winding 26. The primary current is assumed to be a low frequency current. The 20 kHz square wave voltage excitation applied at the excitation winding 24 also causes an additional current to flow in the excitation winding 24. Approximately no current flows in the sense winding 22 given the high impedance input characteristics of the flux sense circuitry 16. The currents in the windings 24 and 26 combine to induce a magnetic flux Φ in the current transformer 14 that is the sum of the magnetic flux Φ1 in the transformer core 20a and the magnetic flux Φ2 in the transformer core 20b:
Φ=Φ₁+Φ₂  (4)

In general, magnetic flux is equal to magnetic flux density integrated over the area through which the flux passes. Therefore, assuming an equal magnetic flux density B1 throughout the transformer core 20a, then the magnetic flux Φ1 is equal to the magnetic flux density B1 multiplied by the effective area A_e1 of the transformer core 20a:
Φ₁=B₁·A_e1  (5)

Likewise, assuming an equal magnetic flux density B2 throughout the transformer core 20b, then the magnetic flux Φ2 is equal to the magnetic flux density B2 multiplied by the effective area A_e2 of the transformer core 20b:
Φ₂=B₂·A_e2  (6)

In general, the voltage across a winding that is wound around a core is equal to the change in magnetic flux in the core with respect to time. The excitation winding 24 is wound around both transformer cores 20a and 20b, and the voltage V_exc across the excitation winding is dependent on the total change in magnetic flux in the core with respect to time: $\begin{array}{cc}{V}_{\mathrm{exc}}=N_{\mathrm{exc}}\frac{d\Phi }{dt}=N_{\mathrm{exc}}\left(\frac{d{\Phi }_1}{dt}+\frac{d{\Phi }_2}{dt}\right)& \left(7\right)\end{array}$

where N_exc is the number of turns of the excitation winding 24. Substituting Eqs. (5) and (6) into Eq. (7) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=N_{\mathrm{exc}}\left(A_{\mathrm{e1}}·\frac{d{B}_1}{dt}+A_{\mathrm{e2}}·\frac{d{B}_2}{dt}\right)& \left(8\right)\end{array}$

As previously noted, the excitation signal applied to the excitation winding 24 may be a 20 kHz square or pulse width modulated voltage signal. For a square wave excitation, the change in magnetic flux density is constant with respect to time (that is, the second derivative of the magnetic flux density is equal to zero). Therefore, the derivative $\frac{d{B}_1}{dt},\frac{d{B}_2}{dt}$
of the magnetic flux density is equal to the total change in magnetic flux density ΔB1, ΔB2 divided by the time interval Δt during which the change occurs: $\begin{array}{cc}\frac{d{B}_1}{dt}=\frac{\Delta B_1}{\Delta t}& \text{(9a)}\\ \frac{d{B}_2}{dt}=\frac{\Delta B_2}{\Delta t}& \text{(9b)}\end{array}$
Substituting Eqs. (9a) and (9b) into Eq. (8) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=N_{\mathrm{exc}}\left(A_{\mathrm{e1}}·\frac{\Delta B_1}{\Delta t}+A_{\mathrm{e2}}·\frac{\Delta B_2}{\Delta t}\right)& \left(10\right)\end{array}$

In general, magnetic flux density is equal to magnetic field intensity multiplied by the permeability of the material. Thus, for the transformer cores 20a and 20b, the magnetic flux densities B₁ and B₂ can be rewritten as follows:
B₁=μ₁·H₁  (11a)
B₂=μ₂·H₂  (11b)

Substituting Eqs. (11a) and (11b) into Eq. (10) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=N_{\mathrm{exc}}\left(A_{\mathrm{e1}}\frac{{\mu }_1·\Delta H_1}{\Delta t}+A_{\mathrm{e2}}\frac{{\mu }_2·\Delta H_2}{\Delta t}\right)& \left(12\right)\end{array}$
Since the excitation winding 24 is wound around both cores 20a and 20b, the change in magnetic field intensity caused by the excitation signal applied to the excitation winding 24 is the same for both the transformer cores 20a and 20b:
ΔH₁=ΔH₂  (13)
Previously, it was assumed that the mean path length of the transformer core 20a is equal to the mean path length of the transformer core 20b (L_e1=L_e2), this arrangement being preferred in order to simplify construction of the transformer 14. If this arrangement is not utilized, then Eq. (12) would also be a function of the ratios of the path lengths $\left(\frac{L_{\mathrm{e1}}}{L_{\mathrm{e2}}}\right).$

Substituting Eq. (13) into Eq. (12) and simplifying yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=N_{\mathrm{exc}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1+A_{\mathrm{e2}}·{\mu }_2\right)& \left(14\right)\end{array}$
where ΔH=ΔH1=ΔH2.

The voltage V_sen across the sense winding 22 can be derived in the same manner as the voltage V_exc across the excitation winding 24, except that the change in the magnetic flux $\frac{d{\Phi }_1}{dt}$
in the transformer core 20a does not contribute to the voltage V_sen across the sense winding 22 because the sense winding 22 is not wound around the transformer core 20a. Thus, the voltage V_sen across the sense winding 22 can be expressed as follows: $\begin{array}{cc}{V}_{\mathrm{sen}}=N_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e2}}·{\mu }_2\right)& \left(15\right)\end{array}$
wherein N_sen is the number of turns of the sense winding 22. Dividing the voltage V_exc across the excitation winding 24 (Eq. 14) by the voltage V_sen across the sense winding 22 (Eq. 15) yields the following relationship: $\begin{array}{cc}\frac{V_{\mathrm{exc}}}{V_{\mathrm{sen}}}=\frac{N_{\mathrm{exc}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1+A_{\mathrm{e2}}·{\mu }_2\right)}{N_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e2}}·{\mu }_2\right)}& \left(16\right)\end{array}$

Canceling and rearranging terms in Eq. (16) yields the following relationship: $\begin{array}{cc}\frac{V_{\mathrm{exc}}}{V_{\mathrm{sen}}}=\frac{N_{\mathrm{exc}}}{N_{\mathrm{sen}}}\left(\frac{A_{\mathrm{e1}}·{\mu }_1}{A_{\mathrm{e2}}·{\mu }_2}+1\right)& \left(17\right)\end{array}$

From Eq. (17), it is seen that the ratio $\frac{V_{\mathrm{exc}}}{V_{\mathrm{sen}}}$
is a function of the parameters N_exc, N_sen, A_e1, A_e2, and the ratio $\frac{{\mu }_1}{{\mu }_2}.$
However, the parameters N_exc, N_sen, A_e1, and A_e2 are constants that relate to the known construction of the current transformer. Further, the voltage V_exc is the amplitude of the 20 kHz square wave excitation and is therefore also known, and the voltage V_sen is measured using the flux sense circuitry 16 (which, as previously noted, comprises a voltage sensing circuit). Therefore, the voltage V_sen is a function of only the ratio $\frac{{\mu }_1}{{\mu }_2}.$

As previously noted, the materials for the transformer cores 20a and 20b are chosen such that the materials have permeabilities both have permeability functions that are a function of flux density, but the permeability functions vary with respect to each other in a way that is not divisible by a single scalar constant. Although both functions vary with respect to flux density, they also vary with respect to each other in a manner that varies in accordance with flux density. Thus, since the ratio $\frac{{\mu }_1}{{\mu }_2}$
is a function of flux density, flux density can be sensed by sensing the voltage V_sen across the sense winding 22.

Referring now to FIG. 5, FIG. 5 is a graph of the magnetic flux in the transformer core 20a as a function of the voltage V_sen (curve 38), the magnetic flux in the transformer core 20b as a function of the voltage V_sen (curve 40), and the total magnetic flux in the both the transformer cores 20a and 20b as a function of the voltage V_sen (curve 42). FIG. 6 is similar, but has magnetic flux density instead of magnetic flux on the Y-axis. Thus, curve 46 is the variation in the magnetic flux density in transformer core 20a as a function of the voltage V_sen, and curve 48 is the variation in the magnetic flux density in transformer core 20b as a function of the voltage V_sen. It is therefore seen that, by storing the information represented in FIG. 5 or FIG. 6 in a look-up table, and accessing the look-up table using the voltage V_sen as an input value, the voltage V_sen can be converted into a flux or flux density value, respectively.

As will become apparent below in connection with FIGS. 10-11, in many applications, an extensive amount of information pertaining to the magnetic flux is not required. For example, it may be adequate to obtain information pertaining simply to whether the flux is above or below a certain value, without there being a need to determine the actual amount of flux. In these applications, the voltage V_sen can be used as indication of the flux directly (e.g., as a feedback signal), without the need for any further processing.

Referring now to FIG. 7, FIG. 7 is a graph of-the magnetic field intensity in the transformer cores 20a and 20b as a function of the voltage V_sen (curve 50). Notably, magnetic field intensity is proportional to current. Therefore, the graph of FIG. 7 can be used to convert the voltage V_sen into a value that represents the primary current. For example, assuming the voltage V_sen is equal to about 77 mV, then the current in the primary winding is equal to about 160 mA. Of course, if the primary current is a low frequency current, then the transformer 14 will tend toward saturation unless measures are taken to prevent saturation from occurring. The use of the voltage V_sen as a feedback signal to prevent saturation is discussed in connection with FIGS. 10-11, below. From FIGS. 5-7, it is also apparent that the voltage V_sen is simultaneously indicative of the magnetic flux, the magnetic flux density, and the magnetic field intensity, regardless of how the voltage V_sen is eventually processed.

Referring now to FIGS. 8 and 9, two graphs showing the operation of the system of FIG. 1 under different operating conditions are shown. In FIG. 8, a 10 Hz square wave excitation is applied to the primary winding 26 and, in FIG. 9, a 500 Hz square winding excitation is instead applied.

In FIG. 8, curve 60 is the envelope of the voltage measured across the sense winding 22. Thus, curve 60a shows the variation of the maximum peaks with respect to time and curve 60b shows the variation of the minimum peaks with respect to time. (Due to the time scale of FIG. 8, which shows a 100 ms interval, it is not possible to show individual cycles of the voltage V_sen, which is a 20 kHz signal.) At any given instant, the voltage V_sen (which, as previously noted, is a peak-to-peak voltage) is the difference between the curve 60a and the curve 60b. The curve 62 is the magnetic flux in the transformer cores 20a and 20b with respect to time. The curve 64 is the 10 Hz square wave voltage excitation signal applied to the primary winding 26. In FIG. 8, one full cycle of the 10 Hz excitation signal is shown.

Beginning at t=0 (in the middle of the graph), it is seen that the magnetic flux Φ is equal to zero. However, due to the positive voltage applied to the primary winding 26, the magnetic flux Φ is increasing. At the same time, the voltage V_sen is at a minimum value (corresponding to the magnetic flux being equal to zero), but is increasing (corresponding to the increasing magnetic flux).

At about 13 ms, the transformer cores 20a and 20b begin to saturate, and therefore the voltage V_sen is no longer a reliable indication of the magnetic flux Φ. In this regard, it may be noted that conventional current transformers and Hall-based current sensing devices similarly no longer produce reliable data once saturation occurs, because the equations that govern “normal operation” of these devices break down in saturation. As will be discussed below, a particular advantage of the system of FIG. 1 is that it detects magnetic flux, which in turn allows the level of magnetic flux to be controlled to prevent saturation.

At about t=35 ms, the 10 Hz square wave excitation changes state such that a negative voltage is now applied to the primary winding 26. Due to the negative voltage, the magnetic flux begins to decrease and, at about t=39 ms, the transformer cores 20a and 20b drop out of saturation. Once this occurs, the voltage V_sen again provides an indication of the magnetic flux of the cores 20a and 20b. The voltage V_sen continues to decrease as a negative voltage is applied to the primary winding 26 and the magnetic flux continues to decrease.

The graph in FIG. 8 ends at t=50 ms, however, it can be seen from the interval during t=50 ms to t=0 that the system operates as expected. Given that the signal at the primary winding 26 is cyclical, the operation of the system during the interval t=−50 ms to t=0 is the same as during the interval t=50 ms to t=100 ms. During these intervals, the voltage V_sen continues to decrease as the negative voltage represented by the curve 64 continues to be applied to the primary winding 26 and the magnetic flux in the transformer cores 20a and 20b continues to decrease.

The transformer cores 20a and 20b are in saturation from t=−38 ms to t=−13 ms. When the transformer cores 20a and 20b come out of saturation, the voltage V_sen is at a maximum value and subsequently decreases as the magnetic flux decreases towards zero.

It may therefore be noted that the voltage V_sen provides an “absolute value” indication of the magnetic flux. In other words, the voltage V_sen is at a positive maximum both when the magnetic flux is at a positive maximum and when the magnetic flux is at a negative maximum, and the voltage V_sen is at a positive minimum when the magnetic flux density is equal to zero (the voltage V_sen does not assume negative values). Thus the voltage V_sen does not indicate the polarity of the magnetic flux density. However, an approach is described below for determining the polarity of the magnetic flux in situations where it is desirable to know the polarity.

In FIG. 9, a 500 Hz square wave excitation is applied to the primary winding 26. The curve 70 is the voltage across the sense winding 22 with respect to time. Because FIG. 9 has a different time scale than FIG. 8 (only 2 ms of data is shown in FIG. 9 as compared to 200 ms in FIG. 8), individual cycles of the voltage V_sen are shown in FIG. 9. The voltage V_sen is the peak-to-peak voltage of the curve 70, the curve 72 is the magnetic flux with respect to time, and the curve 74 is the 500 Hz square wave voltage excitation signal applied to the primary winding 26.

From t=−2.0 ms to t=−1.0 ms (not shown) a negative voltage is applied to the primary winding 26. Accordingly, at t=−1.0 ms, the magnetic flux in the cores 20a and 20b is at a negative maximum, and the voltage V_sef is at a maximum value. Thereafter, the voltage applied to the primary winding 26 becomes positive, and the magnetic flux begins to increase towards zero. At the same time, the voltage V_5e, begins to decrease, and continues to decrease until the magnetic flux passes through zero at the t=−0.5 ms, at which time the voltage V_sen begins to increase.

At t=0, the 500 Hz square wave excitation changes state such that a negative voltage is now applied to the primary winding 26. As a result, the magnetic flux decreases, and the voltage V_sen decreases and then increases after the magnetic flux passes through zero, as expected.

In FIG. 9, a number of additional features can be seen as a result of the different time scale as compared to FIG. 8. First, ripples can be seen in the curve 72 that represents the magnetic flux. These ripples are caused by the 20 kHz square wave excitation applied to the excitation winding 24. In other words, the positive voltage pulses applied to the excitation winding 24 cause slight increases in the magnetic flux ( which cancel the slight decreases caused by the negative voltage pulses (i.e., assuming the 20 kHz excitation signal is centered about zero).

It may be noted, however, that the shape of the pulses will be asymmetric depending on whether the flux in the cores 20a and 20b is positive or negative. In other words, for example, the positive flux excursions will be different depending on whether the flux in the core is positive or negative, because in one situation the core is moving towards saturation (and permeability is decreasing), and in the other situation the core is moving away from saturation (and permeability is increasing). By examining these flux excursions, an indication of the polarity of the flux may be obtained. It may also be that using a lower frequency square wave excitation signal and/or a larger amplitude excitation will increase the size of the flux excursions and therefore make the asymmetry more pronounced. Nevertheless, it is desirable to have the flux excursions as small as possible while still allowing the polarity to be determined, in order to minimize the effect of the excitation signal applied at the winding 24 on the overall amount of flux in the transformer 14.

Additionally, in FIG. 9, it is seen that the average value of the curve 70 changes as a function of the state of the 500 Hz square wave excitation. This effect is also present in FIG. 1 although it is not as noticeable. This effect is caused by the fact that the 500 Hz square wave excitation causes a change in flux with respect to time that is detectable by the sense winding 22. The change in flux caused by the 500 Hz square wave excitation manifests itself as a voltage that is added to or subtracted from the higher frequency voltage oscillation (the peak-to-peak value of which is measured to determine flux) caused by the 20 kHz square wave excitation applied to the excitation winding 24.

The practical consequence of this latter feature is that the frequency difference between the highest anticipated frequency of the signal applied to the primary winding 26 and the frequency of the excitation signal applied to the excitation winding 24 should be sufficiently large to permit the portion of the output of the sense winding 22 attributable to the signal applied to the primary winding 26 and the portion of the output attributable to the signal applied to the excitation winding 24 to be distinguished.

Referring now to FIG. 10, a preferred current sensing system 100 is illustrated. The system 100 of FIG. 10 is the same as the system 10 of FIG. 1, except that in FIG. 10, a summing element 102, a PID (proportional-integral-derivative) gain element 104, a controlled current source 106, a flux cancellation winding 108, burden resistors 112 and 114, and a summing element 116 are also utilized. These additional elements are used to implement a feedback control loop to control the flux in the cores 20a and 20b, and to measure the current through the primary winding 26. In practice, the flux sense circuitry 16, the summing element 102 and the PID gain element 104 may be implemented using analog circuit elements or using digital circuit elements, for example, an AID converter (to measure the voltage V_sen), a microprocessor coupled to a memory (to linearize and scale the output from the A/D converter, and to implement the summing element 102 and the PID gain element 104), and a D/A converter (to provide an analog output to the controlled current source 106). The summing element 116 may be implemented using, for example, analog op-amp circuitry or additional digital circuitry.

The system operates in the following manner. When a current flows in the primary winding 26, the current induces a magnetic flux which is detectable by the flux sense circuitry 16 in the manner previously described in connection with FIG. 1. The flux sense circuitry 16 preferably provides an output which is a linearized version of the voltage V_sen, and provides an output with a value equal to zero when the magnetic flux is equal to zero. This can be achieved through the use of a look-up table, for example.

The output of the flux sense circuitry 16 is provided as feedback to the summing element 102. The summing element 102 also accepts a setpoint input. In FIG. 10, the setpoint is set equal to zero, although it would also be possible to use a different (non-zero) setpoint. This may done, for example, if is desired to maintain the flux at some other, non-zero level.

The output of the summing element 102 is a flux error signal which is applied to the PID gain element 104, which performs PID compensation. Although a PED control loop is shown, other control loops could also be utilized (e.g., a PI control loop or a hysteresis control loop). The compensated error signal from the PID gain element 104 is provided as a control input to the controlled current source 106. Thus, the current source 106 is controlled so as to maintain zero magnetic flux in the transformer cores 20a and 20b. To the extent that there is non-zero flux, this is detected as an error by the flux sense circuitry 116 and the summing element 102, and the PED gain element 104 performs compensation to eliminate the error (i.e., drive the flux to zero).

Advantageously, the system 10 not only takes prevents low frequency current components in the primary winding 26 from saturating the transformer 14, but also makes it possible to measure those low frequency components. The primary current, including both the low frequency and high frequency current components, is measured in the following manner.

With respect to the low frequency components of the primary current, the burden resistor 112 is provided which produces a voltage V₀(dc) that is indicative of the low frequency current components. The current flowing through the burden resistor 112 is a mirror image of the low frequency current flowing through the primary winding 26. This is because the flux sense circuitry 16 utilizes a high impedance voltage measuring device, so no significant current flows through the sense winding 22 and therefore the sense winding 22 does not affect the flux in the transformer cores. Additionally, the current that flows in the winding 24 does not have any low frequency components. Therefore, in order for the flux produced by the current flowing in the winding 108 to cancel the flux produced by the primary current, the current flowing in the winding 108 must be a mirror image of the primary current (related by the turns ratio of the windings 108 and 26). As a result, the low frequency components of the primary current can be measured by measuring the voltage V₀(dc) across the burden resistor 112.

With respect to the high frequency components of the primary current, the burden resistor 114 is provided which produces a voltage V₀(ac) that is indicative of the high frequency current components. For high frequency components of the primary current, the primary winding 26 and the winding 24 behave as a conventional current transformer. The high frequency components of the primary current induce a secondary current in the winding 24, thereby producing the voltage V₀(ac), which is indicative of the high frequency current components. The winding 24 acts as an inductor to filter out the 20 kHz excitation signal in the measured voltage V₀(ac).

Thus, in combination, the high frequency and low frequency components of the system 100 cooperate in the following manner. The flow of primary current in the primary winding 26 induces a magnetic flux that flows in the transformer 14. The natural inclination is for the magnetic flux to induce a secondary current in the winding 24. To the extent that this occurs, the flow of secondary current in the winding 24 produces a counterflux that approximately cancels the flux produced by the primary current. The cancellation of the flux produced by the primary current is not total due to non-zero winding resistance. However, to the extent that high frequency components of the flux are not immediately canceled, due to the non-zero winding resistance, the high frequency components are eventually canceled by virtue of the fact that the primary current is cyclical and therefore the remaining flux is canceled during the next half-cycle.

For conventional current transformers, it is when the frequency is too low, and therefore magnetic flux is allowed to build up for too long before being canceled during the next half cycle, that saturation problems are encountered. In the current sensing system 100 of FIG. 10, this magnetic flux is instead canceled by the current that flows through the winding 108. This current is produced by the controlled current source 106 in response to a control signal from the PID gain element 104. The current required to cancel the remaining flux is measurable by measuring the voltage V₀(dc) across the burden resistor 112. The voltages V₀(dc) and V₀(ac) produced across the burden resistors 112 and 114 can then be directly added by a summing element 116 to produce an overall indication of the primary current.

It may be noted that, in the current sensing system of FIG. 10, there is no particular need to store information that relates to μ₁(H), μ₂(H), μ₁(B), or μ₂(B). The voltage V_sen provides an indication of the magnetic flux in the transformer cores (i.e., by increasing as the flux error increases), but there is no need to determine the actual amount of flux in the transformer cores 20a and 20b. Thus, this further simplifies implementation of the system of FIG. 10.

Referring now to FIG. 11, another embodiment of a current sensing system 120 is illustrated. The embodiment of FIG. 11 is preferred over the embodiment of FIG. 10, because it is simpler. In particular, in FIG. 11, 20 kHz square wave excitation source 18 and the current source 106 have been replaced with a single PWM voltage source 126, the windings 24 and 108 have been replaced by a single winding 122, the burden resistors 112 and 114 have been replaced by a single burden resistor 124.

The voltage source 126 produces a 20 kHz PWM signal which operates in the same manner as the signal from the 20 kHz square wave excitation source 18 in FIGS. 1 and 10. In this case, however, the control signal from the PID gain element 104 controls the pulse widths of the signal from the source 126 to provide the signal with the DC component required to cancel any flux in the transformer 14.

The current flowing through the winding 122 is then a mirror image of the primary current, and is related to the primary current by the turns ratio of the winding 122 to the primary winding 26. The flux induced by the primary current is canceled partially by the induced counterflux from the high frequency components of the current in the winding 122 (the high frequency components of the current in winding 122 being produced in direct response to the primary current via the magnetic coupling of the transformer 14), and partially by the flux induced by the low frequency components of the current in the winding 122 (the low frequency components of the current in winding 122 being produced by the source 26 in response to the control signal from the PID gain element 104). The mid-frequency components of the flux are canceled by a combination of these actions. The net effect is that all of the magnetic flux in the transformer 14 is canceled, and the voltage V₀ produced across the burden resistor 124 which is indicative of all frequency components of the primary current.

Referring now to FIG. 12, another preferred current sensing system 150 is illustrated. The system 150 of FIG. 12 is the same as the system 10 of FIG. 1 except that, in the case of FIG. 12, the sense winding 22 is in practice implemented as two separate sense windings 22a and 22b. Also, in the embodiment of FIG. 12, the flux sense circuitry 16 is shown as comprising a first absolute value circuit 154, a second absolute value circuit 156, and a comparison circuit 158. Further, the system 150 also includes a ±1 gain circuit 162, a controller 164, summing circuits 166 and 168, and a burden resistor 170. Again, the circuitry of FIG. 12 may be implemented using either analog or digital components or a combination of both.

As shown in FIG. 12, the sense winding 22a is wound around the transformer core 20a and not the transformer core 20b, the sense winding 22b is wound around the transformer core 20b and not the transformer core 20a, and the excitation winding 24 and the primary winding 26 are wound around both transformer cores 20a and 20b. This arrangement may be fabricated in a manner similar to that described above in connection with FIGS. 2A-2B. In the case of FIG. 12, however, the transformer core 20a is wound with the sense winding 22a before being placed adjacent to the transformer core 20b (which is wound with the sense winding 22b). After the transformer cores 20a and 20b have been wound with the sense windings 20a and 20b, respectively, the combination of these structures is then wound with the excitation winding 24 and the primary winding 26.

In the embodiment of FIG. 12, the transformer core 20a is Supermalloy provided in the form of a tape wound core. This material can be purchased as Part #47 P 8602 from Magnetic Metals Corporation, 2475 LaPalma Avenue, Anaheim, Calif., 92801. The transformer core 20b may for example be formed of SuperPerm 49 provided in the form of a tape wound core. This material can be purchased as Part #47 P 4902 from the same company. Supermalloy is approximately 80% nickel in composition and SuperPerm49 is approximately 50% nickel by composition. Consistent with what has been previously described, these two materials have different permeability characteristics. It has been found that these materials yield a more linear transfer function and improved signal gain as compared to at least some other materials, although these materials are merely provided as preferred examples and other materials could also be used.

In the embodiment of FIG. 12, the high frequency excitation source 18 is implemented as a sinusoidal source rather than a squarewave source (although other excitation sources could also be used). As previously indicated, the above discussion in connection with Eqs. (10)-(17) was provided in the context of a square wave excitation. For a sinusoidal excitation, Eqs. (10)-(17) are written slightly differently. The following derivation is similar to the derivation provided above in connection with Eqs. (10)-(17), except that it is performed assuming sinusoidal excitation, it is performed for both sense windings 22a and 22b, and it includes a few additional equations that describe one preferred way of comparing the voltages measured from the sense windings 22a and 22b. As will become apparent, it is not necessary that any of the following equations actually be utilized or implemented in a circuit in order for the circuit to take advantage of the principals described herein. Rather, the following discussion merely provides an understanding why signal measurements acquired from the sense windings 22a and 22b provide an indication of the relative permeability of the transformer cores 20a and 20b, and thereby also provide an indication of the magnetic flux (and, inherently, at the same time, magnetic flux density and magnetic field intensity) in the transformer cores 20a and 20b.

Particularly, for a sinusoidal excitation, the voltage across the excitation winding is governed by the following equation: $\begin{array}{cc}{V}_{\mathrm{exc}}=\frac{\pi }{2}{N}_{\mathrm{exc}}\left(A_{\mathrm{e1}}·\frac{\Delta B_1}{\Delta t}+A_{\mathrm{e2}}·\frac{\Delta B_2}{\Delta t}\right)& {\left(10\right)}^{\prime }\end{array}$
where V_exc represents the peak voltage of the sine wave, ΔB is equal to the peak flux density, ΔH is equal to the peak magnetic field intensity, and Δt is equal to the difference in the time from when H is equal to its peak value to the time when H is equal to zero. Rewriting Eq. (10)′ in terms of magnetic field intensity in a manner described above in connection with Eqs. (10)-(12) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=\frac{\pi }{2}{N}_{\mathrm{exc}}\left(A_{\mathrm{e1}}\frac{{\mu }_1·\Delta H_1}{\Delta t}+A_{\mathrm{e2}}\frac{{\mu }_2·\Delta H_2}{\Delta t}\right)& {\left(12\right)}^{\prime }\end{array}$
Rewriting Eq. (12)′ by assuming that ΔH₁=ΔH₂ and L_e1=L_e2 in the manner described above in connection with Eqs. (13)-(14) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{exc}}=\frac{\pi }{2}{N}_{\mathrm{exc}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1+A_{\mathrm{e2}}·{\mu }_2\right)& {\left(14\right)}^{\prime }\end{array}$
where ΔH=ΔH1=ΔH2 and N_exc is the number of turns of the excitation winding 24 (for example, N_exc=1000 in one preferred embodiment).

The voltages V_sen—a and V_sen—b across the sense windings 22a and 22b, respectively, can be expressed as follows: $\begin{array}{cc}{V}_{\mathrm{sen\_a}}=\frac{\pi }{2}{N}_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1\right)& {\text{(15a)}}^{\prime }\end{array}$ $\begin{array}{cc}{V}_{\mathrm{sen\_b}}=\frac{\pi }{2}{N}_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e2}}·{\mu }_2\right)& {\text{(15b)}}^{\prime }\end{array}$
wherein N_sen is the number of turns of the sense windings 22a and 22b and is assumed to be the same for both windings (e.g., N_sen=200). Dividing the voltage V_exc across the excitation winding 24 (Eq. 14)′ by the voltages V_sen—a and V_sen—b across the sense winding 22a and 22b (Eqs. (15a)′-15b′)), respectively, yields the following relationship: $\begin{array}{cc}\frac{V_{\mathrm{exc}}}{V_{\mathrm{sen\_a}}}=\frac{N_{\mathrm{exc}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1+A_{\mathrm{e2}}·{\mu }_2\right)}{N_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1\right)}& {\text{(16a)}}^{\prime }\\ \frac{V_{\mathrm{exc}}}{V_{\mathrm{sen\_b}}}=\frac{N_{\mathrm{exc}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e1}}·{\mu }_1+A_{\mathrm{e2}}·{\mu }_2\right)}{N_{\mathrm{sen}}\frac{\Delta H}{\Delta t}\left(A_{\mathrm{e2}}·{\mu }_2\right)}& {\text{(16b)}}^{\prime }\end{array}$

Canceling and rearranging terms in Eqs. (16a)′-(16b)′ yields the following relationships: $\begin{array}{cc}\frac{V_{\mathrm{exc}}}{V_{\mathrm{sen\_a}}}=\frac{N_{\mathrm{exc}}}{N_{\mathrm{sen}}}\left(\frac{A_{\mathrm{e2}}·{\mu }_2}{A_{\mathrm{e1}}·{\mu }_1}+1\right)& \text{(17a)}’\\ \frac{V_{\mathrm{exc}}}{V_{\mathrm{sen\_b}}}=\frac{N_{\mathrm{exc}}}{N_{\mathrm{sen}}}\left(\frac{A_{\mathrm{e1}}·{\mu }_1}{A_{\mathrm{e2}}·{\mu }_2}+1\right)& \text{(17b)}’\end{array}$

Solving for the voltages V_{sen—hd a} and V_sen—b yields the following relationships: $\begin{array}{cc}{V}_{\mathrm{sen\_a}}=\frac{V_{\mathrm{exc}}}{\frac{N_{\mathrm{exc}}}{N_{\mathrm{sen}}}\left(\frac{A_{\mathrm{e2}}·{\mu }_2}{A_{\mathrm{e1}}·{\mu }_1}+1\right)}& \text{(18a)}\\ V_{\mathrm{sen\_b}}=\frac{V_{\mathrm{exc}}}{\frac{N_{\mathrm{exc}}}{N_{\mathrm{sen}}}\left(\frac{A_{\mathrm{e1}}·{\mu }_1}{A_{\mathrm{e2}}·{\mu }_2}+1\right)}& \text{(18b)}\end{array}$

Subtracting the voltage V_sen—a from the voltage V_sen—b and assuming the transformer cores 20a and 20b are constructed so as to have equal cross sections (i.e., A_e1=A_e2) yields the following relationship: $\begin{array}{cc}{V}_{\mathrm{sen\_b}}-V_{\mathrm{sen\_a}}=\frac{N_{\mathrm{sen}}}{N_{\mathrm{exc}}}{V}_{\mathrm{exc}}\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1}& \left(19\right)\end{array}$

From Eq. (19), it is seen that the differential term V_sen—b−V_sen—a is a function of N_exc, N_sen, V_exc, and the ratio $\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1}.$
However, the parameters N_exc and N_sen, are constants that relate to the known construction of the current transformer. Further, the voltage V_exc is the amplitude of the sinusoidal excitation and is therefore also known. Therefore, the differential term V_sen—b−V_sen—a is a function of only the ratio $\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1},$
which varies in accordance with the magnetic flux (or magnetic flux density or magnetic field intensity) in the transformer cores 20a and 20b.

As compared to Eq. (17), discussed above and in which the voltage V_sen is a function of the ratio $\frac{{\mu }_1}{{\mu }_2},$
the differential term V_sen—b−V_sen—a in Eq. (19) is a function of $\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1}.$
This is advantageous inasmuch as non-linearities associated with the permeability termed to cancel, making the term $\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1}$
a more linear function of the amount of flux in the transformer cores 20a and 20b. Additionally, the differential term V_sen—b−V_sen—a also improves signal gain (that is, the differential term V_sen—b−V_sen—a experiences greater variations as a function of flux in the transformer cores 20a and 20b) and thereby yields an improved signal-to-noise ratio.

In the most preferred embodiment, rather than use the differential term V_sen—b−V_sen—a an indication of flux, the average value of the difference in absolute values of the voltages V_sen—a and V_sen—b is employed: $\begin{array}{cc}\mathrm{AVG}\left[V_{\mathrm{sen\_b}}-V_{\mathrm{sen\_a}}\right]=\frac{2}{\pi }\frac{N_{\mathrm{sen}}}{N_{\mathrm{exc}}}{V}_{\mathrm{exc}}\frac{{\mu }_2-{\mu }_1}{{\mu }_2+{\mu }_1}& \left(20\right)\end{array}$
where AVG[ ] is an averaging operator and | | is an absolute value operator. For a low frequency signal applied to the primary winding 26, the term AVG[|V_sen—b|−|V_sen—a|] will have a low frequency and will have a value that increases as the amount of magnetic flux in the transformer cores 20a and 20b increases.

As previously mentioned, it is not necessary that any of the foregoing equations actually be computed in a circuit in order for the circuit to take advantage of the above phenomenon. Rather, the above discussion is provided merely for purposes of demonstrating that signal measurements (in this case, voltage measurements of the voltages V_sen—a and V_sen—b) may taken across the sense windings 22a and 22b and used to provide an indication of the magnetic flux (and, inherently, at the same time, magnetic flux density and magnetic field intensity) in the transformer cores 20a and 20b. In the currently preferred embodiment, two voltage measurements V_sen—a and V_sen—b are taken across the sense windings 22a and 22b, processed to produce the differential term AVG[|V_sen—b|−|V_sen—a|],and this differential term is then used as an indication of flux in the transformer cores 20a and 20b. It will be appreciated that this is just one way of processing and using voltage measurement(s) from the sense winding(s), and that many other configurations may be employed for acquiring information pertaining to the relative permeability of the transformer cores 20a and 20b.

The operation of the system 150 will now be described in greater detail. In operation, when a current flows in the primary winding 26, the current induces a magnetic flux which is detectable by the flux sense circuitry 16 in the manner previously described in connection with FIG. 1. The flux sense circuitry 16 includes the absolute value circuits 154 and 156 and the comparison circuit 158 which produce the differential term AVG[|V_sen—b|−|V_sen—a|]. The comparison circuit 158, which in this case is a subtraction circuit, is also assumed to be constructed so as to include an implicit averaging function. The flux sense circuitry 16 is preferably formed of operational amplifiers or other high input impedance devices to avoid the sense windings 22a and 22b affecting the amount of flux in the transformer cores 22a and 22b as previously described. The comparison circuit 158 in this embodiment is a difference circuit that computes the difference between the absolute values of the voltages V_sen—a and V_sen—b.

In general, the controller 164 then uses the output of the comparison circuit 158 to control the current through the excitation winding 24 such that the magnetic flux produced by the excitation winding 24 cancels the magnetic flux produced by the primary winding 22. In practice, as will be described below, the controller 164 has two stable operating points at which the absolute value of the flux is slightly greater than zero. To the extent that the flux is not at one of these two operating points, this is detected as an error by the flux sense circuitry 16, and the controller 164 performs compensation to eliminate the error. Although the controller 164 operates to maintain magnetic flux near a zero setpoint in this embodiment, it will be appreciated that other fixed or varying setpoints could also be used depending on the application.

From a control loop approximation standpoint, for signal components from DC to approximately 2 Hz, the controller 164 zeros flux in the transformer cores 20a and 20b based on the error signal provided by the flux sense circuitry 16. For signal components above 2 Hz, the controller operates on the derivative of flux (or voltage), and zeros the voltage imposed on the transformer cores 20a and 20b based on the signal provided by the summing element 166. In one particular embodiment, the output of the summing element 166 is coupled to the positive input terminal of an operational amplifier of the controller 164, with the negative input terminal of the operational amplifier being coupled to ground and the burden resistor 170 being coupled in a negative feedback configuration to the output terminal of the operational amplifier and ground. This thereby results in more accurate flux cancellation, allows the flux to be maintained more closely at zero, and allows smaller transformer cores 20a and 20b to be used. The voltage feedback circuit is described in greater detail below in connection with FIGS. 16A-16B. Two feedback paths are thus developed in the system 150, a flux feedback path and a voltage feedback path. For signal components at higher frequencies (e.g., 10 kHz or more), the system 150 operates in a manner similar to a conventional current transformer.

Referring now also to FIG. 13, FIG. 13 is a graph showing the output of the circuit 158. Curve 180 represents actual error signal characteristics with hysteresis and curve 182 is a linearized version of curve 180. As shown in FIG. 13, as the magnitude of the magnetic flux in the transformer cores 20a and 20b increases, the output of the circuit 158 increases. However, it may be noted that positive flux and negative flux yield the same value at the output of the circuit 158, and therefore it is difficult to determine whether the system is operating in the first or second quadrant. Consequently, the +1 gain amplifier 162 is used to overcome this problem. The +1 gain amplifier 162 may for example be controlled by a microprocessor that implements a comparator and flip-flop function. When the correct polarity is selected for feedback in a given quadrant, the feedback is negative and this provides stable control at one of the two stable operating points 186 and 188. When the flux moves from one quadrant to the other (e.g., due to a changing signal at the primary winding 26), the selected polarity is no longer correct, and the resulting feedback becomes positive and this provides unstable control. The flux in the transformer cores 20a and 20b then increases in due to positive feedback. At some point during this ramp, the comparator level is exceeded and the flip-flop is toggled. This in turn inverts the flux feedback from its previous state and a negative feedback is established, resulting in stable control at the other of the two operating points 186 and 188. In this manner, the +1 gain amplifier thus monitors stability of the controller 164 and adjusts operation of the controller 164 upon detecting that the direction of the magnetic flux produced in the transformer cores 20a and 20b should be reversed in order to avoid saturation.

As also shown in FIG. 13 is a negative offset 184 which is preferably introduced into the output of the circuit 185 in order to enhance the stability of the system 150. Without such an offset, it is possible that the magnetic error feedback transfer function Y-axis minimum (that is, the minimum value of AVG[|V_sen—b|−|V_sen—a|]) may have a positive value due to circuit imperfections. If the controller 164 has an integral term for eliminating steady state error as is normally desirable, and if the magnetic error feedback transfer function has a Y-axis minimum that is greater than zero, then the controller 164 will not find a steady state operating point. In this scenario, there is not an operating point at which the error is equal to zero. Therefore, the error feedback transfer function minimum should have a Y-axis value of less than zero as in FIG. 13, allowing the controller 164 to find a steady state operating point. In FIG. 13, two steady state operating points exist at points 186 and 188. The offset 184 maybe introduced, for example, by the circuit 158 or by the controller 164.

Another effect of making the Y-axis minimum of the error feedback transfer function less than zero is that it produces an output offset in the output voltage V_o of the system 150. This offset may be compensated by providing a compensation circuit in the controller 164 that provides additional current for the resistor 170. The polarity of the output of the +1 gain amplifier 162 may be used to determine the polarity of the current needed to provide the proper offset. The current through the burden resistor 170, and thereby the output voltage V_o across the burden resistor 170, may thereby be adjusted to compensate operation of the transfer function offset circuit. Magnetic hysteresis of the materials that form the transformer cores 20a and 20b can be zeroed in a similar fashion. Because the toggle point of the t1 gain amplifier 162represents a certain flux level, the “history” of flux is known. When ±1 gain amplifier 162 toggles and negative feedback is restored, it is possible to sum in a current into the burden resistor 170 of the correct amount to compensate for hysteresis offset.

Other than in the respects described above, the operation of the system 150 is generally the same as the systems 100 and 120. The controller 164 controls the current through the winding 24 such that the flux produced by the winding 24 offsets the flux produced by the winding 22. The flux sense circuitry 16 preferably utilizes a high impedance voltage circuits such as operational amplifiers, so no significant current flows through the sense windings 22a and 22b and therefore the sense winding 22a and 22b do not affect the flux in the transformer cores. Therefore, the current flowing through the excitation winding 24 and the burden resistor 170 is a mirror image of the current flowing through the primary winding 26 (related by the turns ratio of the windings 24 and 26). For high frequency components of the primary current, the primary winding 26 and the winding 24 behave as a conventional current transformer. As a result, the voltage Vo across the burden resistor 170 is proportional to the current through the primary winding 26.

FIG. 14 is a detailed schematic diagram of one possible implementation of the system 150 shown in FIG. 12. Component values for the components in the circuit of FIG. 14 are as follows:

Component Value

C9        1500 pF
C15       .22 uF
C16       4700 pF
C17       4700 pF
C24       1000 pF
C25       .01 uF
D12       D1N4148 MA3Z793
D3        D1N4148 MA3Z793
D4        D1N4148 MA3Z793
D5        D1N4148 MA3Z793
D6        D1N4148 MA3Z793
D7        D1N4733
D8        D1N4733
J1        J174
L2        1000
L4        1000
L5        200
L6        200
M3        2N70027ZTX
M7        2N70027ZTX
Q4        BCX55-16 MPS6714
Q5        BCX52-16 MPS6726
R1        50 Ω
R20       2 KΩ
R21       1 KΩ
R22       15.83 KΩ
R27       10 KΩ
R30       10 KΩ
R31       20 KΩ
R32       10 KΩ
R35       10 KΩ
R37       10 KΩ
R38       10 KΩ
R39       20 KΩ
R40       10 KΩ
R41       10 KΩ
R42       20 KΩ
R43       10 KΩ
R44       10 KΩ
R45       20 KΩ
R46       1.65 KΩ
R47       1.65 KΩ
R48       1 KΩ
R49       800 Ω
R50       150 Ω
R51       2 KΩ
R52       1 KΩ
R53       10 Ω
R54       4.99 KΩ
R55       4.99 KΩ
R56       15.8 KΩ
R60       2 KΩ
R61       2 KΩ
R68       10 KΩ
R69       16.2 KΩ
R70       5.36 KΩ
R71       8.06 KΩ
R72       20 KΩ
R74       4.99 KΩ
R75       8.05 KΩ
R77       1 KΩ
R79       187 KΩ
R80       374 KΩ
U9        OPA4132
U10       OPA4132
U11       OPA4132
U12       OPA4132
U16       OPA4132
U17       OPA4132
U18       OPA4132
U19       OPA4132
U26       ATINY_15L

FIG. 15 is a high level block diagram showing the feedback control loops of FIGS. 10-12. As shown in FIG. 15, each of the circuits shown in FIGS. 10-12 is capable of receiving a flux setpoint (typically, at or near zero) and using the flux set point to control the amount of magnetic flux in a magnetically conductive structure, such as the structure formed by the transformer cores 20a and 20b. The flux sense circuitry 16 implements a relative permeability sensor to acquire information pertaining to the permeability of the transformer core 20a relative to the transformer core 20b. This information is used as feedback to control the amount of flux in the transformer cores 20a and 20b.

FIGS. 16A and 16B show a feedback control circuit comprising a voltage regulator circuit that may be used to control magnetic flux in a transformer core. FIG. 16A shows an embodiment in which such a circuit is implemented independently of a relative permeability-based system. Thus, in FIG. 16A the system comprises a magnetically conductive structure 210 (such as a transformer core) and a feedback control circuit. The feedback control circuit comprises an excitation winding 211, a sense winding 214; a resistor 215, an amplifier 217, and a capacitor 219. The windings 214 and 215 are wound around the magnetically conductive structure 210. The feedback control circuit acquires feedback information pertaining to a voltage across the sense winding 214 and provides an excitation signal to the excitation winding to control the amount of flux in the magnetically conductive structure 210.

In operation, magnetic flux is introduced into the magnetically conductive structure 210. For example, the system may comprise a primary winding 213 which is wound around the magnetically conductive structure 210 and introduces magnetically flux when a current flows through the primary winding. In this embodiment, it is assumed that the current is a non-DC current, comprising for example components at 2 Hz to 10 kHz. The change in flux in the structure 210 caused by the primary current is sensed as a voltage $\left(V\propto \frac{d\Phi }{dt}\right)$
at the sense winding 214.

The resistor 215, amplifier 217, capacitor 219 of the feedback control circuit implement an integrator thereby allowing linear (proportional-integral) feedback control to be achieved. The resistor 215, amplifier 217, capacitor 219 are coupled between the sense winding 214 and the excitation winding 211 and receive the information pertaining to the voltage across the sense winding 214 as an error signal. The error signal is compared with a flux setpoint signal. In the illustrated embodiment, the amplifier 217 is an operational amplifier that receives the error signal at its negative input terminal and has its positive input terminal coupled to common. The flux setpoint signal in this embodiment is therefore a zero setpoint signal. The feedback control circuit is operative to control flux in the magnetically conductive structure in accordance with the flux setpoint signal. The presence of an active source such as the amplifier 217 allows a current to be developed in the excitation winding 24 that more accurately cancels the flux produced by the current in the primary winding 26.

The feedback control circuit is coupled to the winding 211 and to a burden resistor 218. The excitation signal provided by the feedback control circuit at the output of the amplifier 231 results in a current that flows through the excitation winding 211 and through the burden resistor 218. The current in the excitation winding 211 cancels the flux produced by the current in the primary winding 213. The flow of the current through the burden resistor 218 produces a voltage that is indicative of a current through the primary winding 213.

FIG. 16B shows how the circuit of FIG. 16A is implemented in the context of the system 150 of FIG. 12. In FIG. 16B, the reference numbers 211, 213, and 218 have been changed to 24, 26, and 170 to show correspondence with the system 150 of FIG. 12. Additionally, the magnetically conductive structure 210 is now two separate magnetically conductive structures 20a and 20b, preferably having different permeability characteristics as previously described. Additionally, the winding 214 is now implemented as the two separate sense windings 22a and 22b. The reference numbers 215, 217, and 219 are unchanged to show the correspondence with the circuit of FIG. 16A. In FIG. 12, these circuit components are implemented as part of the controller 164. The output of the amplifier 217 also passes through the summation circuit 168, which is preferably referenced to common and therefore does not affect operation of the excitation signal produced by the amplifier 217. The circuit of FIG. 16B operates in generally the same manner as the circuit of FIG. 16A, except that the summation circuit 166 is used to sum the voltages sensed by the sense windings 22a and 22b, which are each wound around only one of the transformer cores 20a and 20b. The circuit 166 therefore sums the change in flux sensed by the sense winding 20a with the change in flux sensed by the sense winding 20b to produce a signal indicative of the total change in flux in the transformer cores 20a and 20b. Thus, it is possible to use the flux control techniques of FIG. 16A either alone or combination with the flux control techniques described above that rely on the relative permeability of multiple magnetically conductive structures.

From the foregoing discussion, a number of advantages of the preferred sensing systems and methods should be apparent. First, the system sensor is able to sense magnetic flux in static conditions, that is, when the change in flux with respect to time is equal to zero. Thus, the system is able to sense and respond to low frequency magnetic flux components including DC.

Additionally, the preferred system is simple in construction. The preferred sensor comprises two magnetic structures or paths formed of materials with different magnetic characteristics, and a plurality of windings. For many applications, such as current sensing, processing can be performed using exclusively analog circuitry, if desired.

Moreover, the preferred flux sensor exhibits superior sensitivity and accuracy characteristics as compared to a Hall-effect sensor. Because it is not necessary to introduce a gap into the magnetic path through which magnetic flux flows, the path retains superior permeability characteristics. Therefore, sensitivity and accuracy remain high. Additionally, if a differential term is used to processed information for the two transformer cores, accuracy and linearity are further improved.

For a current sensor in accordance with the preferred embodiments of the invention, the current sensor operates as a current transformer for high frequency components of the primary current, and therefore retains the favorable sensitivity and accuracy characteristics of conventional current transformers. However, the current sensor is also able to detect low frequency components of the primary current, cancel the flux produced thereby to keep the transformer out of saturation, and generate information that is indicative of the low frequency current components. Thus, the preferred current sensor enjoys the advantages of conventional current transformers in terms of precision and accuracy, but does not suffer the limitations of conventional current transformers (namely, the inability of conventional current transformers to operate in the presence of a primary current with low frequency components).

Many other changes and modifications may be made to the present invention without departing from the spirit thereof. The scope of these and other changes will become apparent from the appended claims.

Claims

1. A current sensor comprising:

a primary winding, wherein the primary winding is wound around a first transformer core and a second transformer core, wherein the second transformer core has a different permeability than the first transformer core and the permeability of the two cores relative to each other varies according to an amount of magnetic flux in the cores;

a sensing winding, wherein the sensing winding is wound around the second transformer core; and

a canceling winding, wherein the canceling winding is wound around the first and second transformer cores.

2. The current sensor according to claim 1, wherein the sense winding is wound around only the second transformer core.

3. The current sensor according to claim 1, further comprising a flux sense circuit coupled to the sense winding wherein voltage across the sensing winding is a function of a ratio of the relative permeabilities of the transformer cores, wherein the ratio of the permeabilities produces an indication of the amount of magnetic flux in the cores; the magnetic flux of the transformer cores being measured by the flux sense circuit.

4. The current sensor according to claim 3, wherein the flux sense circuit is configured to determine the magnetic flux of the transformer cores by reference to a look-up table.

5. The current sensor according to claim 1, wherein a canceling current in the canceling winding mirrors a primary current of the primary winding, wherein the canceling current mirrors the primary current and cancels the magnetic flux in the cores created by the primary current, thus preventing saturation of the transformer cores.

6. A current sensor comprising:

a first magnetically conductive structure and a second magnetically conductive structure, wherein the first and second magnetically conductive structures have different relative permeabilities;

a primary winding, a canceling winding and an excitation winding, wherein the primary winding, the canceling winding, and the excitation winding are wound around the first and second magnetically conductive structures;

a sense winding, wherein the sense winding is wound around only one of the magnetically conductive structures;

a feedback control circuit configured to control a controlled current source, wherein the feedback control circuit comprises a flux sensing circuit, a summing circuit, and a pid circuit, wherein the flux sensing circuit receives a signal from the sense winding, the signal being a function of a ratio of the relative permeabilities of the magnetically conductive structures, and the flux sensing circuit outputs a signal indicative of the magnetic flux in the magnetically conductive structures; and,

a high frequency excitation source, wherein the high frequency excitation source produces a relatively high frequency current compared to the current in the primary winding and is coupled with the excitation winding.

7. The current according claim 6, wherein in operation the summing circuit sums the output signal from the flux sensing circuit with a setpoint input; the summing circuit producing a flux error signal, wherein the pid circuit receives the flux error signal and outputs a compensated error signal, wherein the controlled current source receives the compensated error signal from the pid circuit and outputs a current to the canceling winding approximately equal to the current in the primary winding; the canceling winding canceling the flux in the magnetically conductive structures and preventing saturation of the magnetically conductive structures.

8. The current sensor according to claim 6, further comprising a low frequency burden resistor and a high frequency burden resistor; the burden resistors each being coupled with the canceling winding and the excitation winding, wherein the sum of voltage across the burden resistors indicates a current in the primary winding.

9. A current sensing system comprising:

a first transformer core and a second transformer core, wherein the first and second transformer cores have different relative permeabilities, and wherein the relative permeability of each transformer core relative to the other transformer core varies as a function of the flux density in the transformer cores;

a primary winding, a canceling winding and a sense winding, wherein the primary winding and the canceling winding are wound around both the first and second transformer cores, and the sense winding only being wound around one transformer core; and,

a feedback control circuit controlling a voltage source, wherein the feedback control circuit is coupled with the sense winding and comprises a flux sensing circuit and a summing circuit.

10. The current sensing system according to claim 9, wherein the flux sensing circuit is coupled with the sense winding and the output signal of the flux sensing circuit indicates the magnetic flux in the transformer cores; wherein the summing circuit receives the output signal from the flux sensing circuit and sums it with a setpoint value to produce a flux error signal representing the flux in the transformer cores.

11. The current sensing system according to claim 9, wherein a pid circuit receives the flux error signal and outputs a compensated error signal; wherein the compensated error signal from the pid circuit controls a pulse width outputted by the voltage source, creating a current in the canceling winding equal to the current in the primary winding, the canceling winding canceling flux in the magnetically conductive structures and thus preventing saturation.

12. The current sensing system according to claim 9, wherein a burden resistor is coupled with the canceling winding and provides a voltage indicative of a current in the primary winding.

13. A current sensing system comprising:

a first transformer core and a second transformer core, wherein the first and second transformer cores have different relative permeabilities;

a primary winding, a plurality of sense windings, and a canceling excitation winding, wherein the primary winding and the canceling excitation winding are wound around the first and second transformer cores, and a first sense winding is wound around only the first transformer core and a second sense winding is only wound around the second transformer core; and,

a flux sensing circuit coupled to the first and second sense windings, wherein the flux sensing circuit comprises a plurality of absolute value circuits and a comparator circuit.

14. The current sensing system according to claim 13, wherein each of the sense windings is coupled with a first summing circuit; the output of the comparator circuit indicating the magnetic flux in the transformer cores.

15. The current sensing system according to claim 13, wherein a second summing circuit sums the output of a controller, with a signal from a sinusoidal oscillator to create a signal which is coupled with the canceling excitation winding; the signal passing through the canceling excitation winding canceling the flux in the transformer cores created by the primary winding and precluding saturation of the transformer cores.

16. The current sensing system according to claim 13, further comprising a burden resistor coupled to the canceling excitation winding, wherein voltage across the burden resistor indicates a current in the primary winding.